290

Polymer(Korea), Vol. 40, No. 2, pp. 290-297 (2016)

http://dx.doi.org/10.7317/pk.2016.40.2.290

ISSN 0379-153X(Print)

ISSN 2234-8077(Online)

상전이 및 가압법에 의한 나노여과 중공사 복합막의 제조 및 이의 성능 연구

권세환 · 임지원†

한남대학교 화공신소재공학과

(2015년 11월 6일 접수, 2015년 12월 31일 수정, 2016년 1월 5일 채택)

Preparation of Hollow Fiber NF Membranes by Phase Separation and

Pressurization (PSP) Method and Their Performance Studies

Se Hwan Kwon and Ji Won Rhim†

Department of New Materials and Chemical Engineering, Hannam University, Daejeon 34054, Korea

(Received November 6, 2015; Revised December 31, 2015; Accepted January 5, 2016)

초록: 염석효과를 이용한 “상전이 및 가압법”으로 복합막을 제조하였다. 수용성고분자인 폴리스티렌설폰산(poly(sty-

rene sulfonic acid; PSSA)과 폴리에틸렌이민(polyethylene imine; PEI)을 코팅소재로 사용하였다. 질산마그네슘의 첨

가에 의해 생성된 고분자 입자들을 다공성 폴리비닐리덴플루오라이드 막의 표면에서 가압하여 기공 내부로 흘러 들

어가게 하여 기공 막힘에 의한 복합막을 여러 실험 조건에서 제조하였다. 제조된 막은 가정용 정수기에 적용 가능

성을 알아보고자 4기압 하에서 100 ppm 소금 염용액에 대해 배제율과 투과도의 항으로 특성파악을 수행하였다.

PSSA를 4분 동안 그리고 PEI를 2분 동안 코팅한 2중층 코팅 복합막은 배제율 81.7% 그리고 투과도 148.3 LMH를

보여주었다.

Abstract: A new technique named “phase separation and pressurization” (PSP) was introduced for the preparation of

composite membranes by utilizing the salting-out effect. The water-soluble polymers such as poly(styrene sulfonic acid)

(PSSA) and polyethylene imine (PEI), were used as the coating materials. Polymer particles formed by the addition of

Mg(NO3)26H2O were pressurized and flown to the surface of microporous polyvinylidene fluoride (PVDF) membrane to

prepare composite membranes by pore blockings under various conditions. The resulting membranes were characterized

in terms of the flux and rejection for 100 ppm NaCl at 4 atm to examine their suitability for application in a household

water purifier. The double layer composite membrane prepared by a first coating of PSSA for 4 min and then a second

coating of PEI for 2 min showed the best performance with flux of 148.3 LMH and rejection rate of 81.7%.

Keywords: hollow fiber membrane, nanofiltration, layer-by-layer, salting-out, phase separation.

Introduction

It is well-known that the nanofiltration (NF) membrane pro-

cess is a type of pressure-driven membrane process between

the reverse osmosis and ultrafiltration membrane processes. Its

use has been increasing gradually worldwide because of the

high retention of multivalent anion salts and organic molecules

with a molecular weight range from 100 to 1000 as well as

other advantages such as low operating pressure, low invest-

ment, and low operation and maintenance costs.1 Furthermore,

NF is a promising technique for production of drinking water

from surface and ground water.2,3

With all these advantages, the development of better NF

membranes has received increasing interest globally. The typ-

ical configuration of an NF membrane is the composite type

obtained by forming a 100-200 nm thick ultra-thin dense layer

on a 50 μm thick porous substrate. Many commercial thin film

composite NF or reverse osmosis (RO) membranes are pre-

pared by interfacial polymerization of a polyamide layer on top

of a porous polysulfone support,4-8 plasma-initiated polym-

erization,9-12 photo-initiated polymerization,13 photo graft-

ing,14,15 electron beam irradiation,16 dip-coating,17 phase

inversion,18 chemical cross-linking,19 or layer-by-layer (LbL)

deposition.20 Among these methods, the LbL technique has

been used to prepare high water flux NF composite mem-

branes.21-23 The properties of LbL membranes can be altered

†To whom correspondence should be addressed.E-mail: jwrhim@gmail.com

©2016 The Polymer Society of Korea. All rights reserved.

Preparation of Hollow Fiber NF Membranes by PSP Method and Their Performance Studies 291

Polymer(Korea), Vol. 40, No. 2, 2016

easily by changing polyelectrolyte type, charge density, ionic

degree, etc., but bulk production is limited because LbL mem-

brane preparation is time consuming. As a result, NF mem-

brane preparation techniques are changing to relatively simple

and low-cost chemical cross-linking methods. Gao et al. 24-27

prepared a series of charged NF membranes of different chi-

tosan derivatives cross-linked by various cross-linking agents.

Xu and Yang28 prepared a positively charged NF membrane

from poly(2,6-dimethyl-1,4-phenylene oxide) by in situ

amines cross-linking. Then researchers of these LbL multilayer

film making methods turned to more convenient preparation

techniques with polyelectrolyte complex (PEC) membranes

where both LbL and PECs have the same ionic cross-linking

structures.29-31

In addition, the dipping method and dipping accompanied by

the salting-out method, which is defined as the separation of an

organic phase from an aqueous phase by the addition of a salt,

could be alternatives for the preparation of NF composite

membranes. Because the solubility of water-soluble polymers

in water is reduced by the addition of the inorganic salts, the

polymers are precipitated owing to the change of solubility. 32

In this way, precipitated organic molecules (polymer particles

by phase separation) in the aqueous solution produced by the

salting-out effect can be easily adsorbed on any membrane sur-

faces32-35 and also be deposited on (or block) the porous struc-

ture, thereby leading to a new technology for the preparation of

NF/RO membranes. McDonald and Ries have invented a

preparation method for composite membranes by depositing a

dispersion of discrete polymer particles in a liquid on a porous

substrate to obtain composite membranes and then annealing

the deposited membranes to stabilize them.36

In this study, a new technique is introduced for the prepa-

ration of composite membranes based on the salting-out effect.

The precipitated polymers by the salting-out effect can be

coated onto porous membrane surfaces by the following two

methods: (i) dipping and (ii) blocking the pore structures and

then deposition by pressurizing the precipitated polymer par-

ticles. In general, the typical fouling phenomena can be

regarded as the blocking of membrane pores and/or forming of

a gel layer by pollutants so that both reduce the intrinsic flux.

The proposed method in this study is similar to the fouling for-

mation process. The coating material (usually water-soluble

polymers or mostly polyelectrolytes) dissolved in water

becomes particulates through addition of the appropriate salts

so that the polymer particles are precipitated, and the process

is controlled by the polymer concentration and the quantity of

salts, that is, the ionic strength. Then the porous membranes to

be coated are soaked in this polymer precipitated coating solu-

tion with vigorous stirring for the desired time or the polymer

precipitated coating solution is pressurized and flown to the

porous membrane surface so as to form a top skin layer by

both blocking the pores and being deposited on the surface.

The resulting membranes are water-soluble; therefore, they

may be subject to post-treatments such as cross-linking or

deposition of a counter-ion, i.e., LbL multilayer, etc. From now

on, we call the former “phase separation and solution (PSS)”

coating method and the latter “phase separation and pres-

surization” (PSP) method. In this study, the PSP method will

be investigated in more detail for the preparation of NF hollow

fiber composite membranes. An opposite ionic polymer is

deposited on the first layer by the PSP method to enable the

formation of a water-resistant LbL skin layer.

The polyelectrolytes poly(styrene sulfonic acid) (PSSA) and

polyethylene imine (PEI) were used as coating materials. To

prepare the composite membranes, the PSP method was used.

The polymer particles formed by the addition of Mg(NO3)26H2O

were pressurized and flown to the surface of a porous PVDF

membrane under varying conditions of polyelectrolyte con-

centration, annealing temperature, coating time, cross-linking

agent concentration, etc. The resulting membranes were char-

acterized in terms of flux and rejection for 100 ppm NaCl at 4

atm to determine their suitability for application in a household

water purifier.37

Experimental

Materials. Poly(styrene sulfonic acid) (PSSA, Mw 10000)

and polyethylene imine (PEI, Mw 750000) were purchased

from Sigma-Aldrich (Milwaukee, USA). The magnesium

nitrate hexahydrate (Mg(NO3)26H2O, 99%) along with isoph-

thaloyl acid (IPC) as the cross-linking agent were also pro-

vided by Sigma-Aldrich (Milwaukee, USA). Ultra-pure water

was provided by Younglin Instrument, Aqua MAX™. All

reagents and solvents were used without further purification.

Preparation of Composite Membrane. The preparation

method is based on the forced fouling induction of polymer

particles precipitated by the salting-out effect. The salt used for

the salting-out effect was Mg(NO3)26H2O33-36 and its ionic

strength (IS) as calculated by eq. (1) ranged from 0.1 to 0.3.

(1)IS1

2--- c

izi

2

i 1=

n

∑=

292 S. H. Kwon and J. W. Rhim

폴리머, 제40권 제2호, 2016년

where ci is the concentration of ion i, zi is the charge of ion i,

and Σ is the sum taken over all ions in the solution.

The concentrations of polymer solutions used for the PSP

methods were in the range of 1% to 4% by weight. First, the

appropriate polymer was dissolved in water and then the

desired IS quantity of Mg(NO3)26H2O was added. Next, a sim-

ple potting module was prepared with 20 cm-long hollow

fibers of polyvinylidene fluoride (PVDF) microfiltration (MF)

membranes supplied from Waters Co. Ltd. (Daejeon, Korea,

average pore size 0.03 m), and this was installed in the PSP

coating apparatus (Figure 1). The polymer precipitated solu-

tion was circulated for the desired time and pressure. Then the

prepared membrane was dried in the air and annealed for the

desired time at the desired temperature. In case an LbL mem-

brane was prepared, another ionic polymer was then deposited

on the first formed layer.33,34 Afterwards, if necessary, the

resulting coated membranes were soaked in a cross-linking

bath for the desired time.

Surface Characterization. The surface morphologies of

the pristine and coated surfaces of the NF PVDF membranes

were analyzed using a field emission scanning electron micro-

scope (FE-SEM, Hitachi S-4800, Tokyo, Japan).

Membrane Performance. Nanofiltration performance was

determined in terms of water flux and salt rejection by a per-

meation apparatus (Figure 2) using a hollow membrane cell

with an effective area of 32.1 cm2. The membrane perfor-

mance determination was carried out at an operating pressure

of 4 atm using a 100 ppm NaCl solution at 25 oC under a con-

stant flow rate of 1.2 L/min. Water flux was determined by

direct measurement of the permeate flow.

Flux [LMH] = Permeate (L)/Membrane area (m2) × Time (h)

(2)

Salt rejection was determined in accordance with the salt

concentration in the permeate solution, which was obtained

through measurement of the electrical conductance of the per-

meate and feed solutions using a conductivity meter (Orion,

Model 162).

Rejection (%) = (3)

Results and Discussion

Surface Analysis. With the PSP coating method, the result-

ing membranes appeared to be well-coated because obser-

vation of the pores showed that the pores were not open. In a

similar manner to the fouling phenomena, with the PSP coat-

ing method the precipitated polymer particles, which are pre-

cipitated after the addition of salt leads to a change in the

solubility of the polymer in the water, block pores of all sizes.

Hence, this PSP coating method does not make a distinction

between ultrafiltration (UF) or MF membranes used as the

support layers, and it is not affected by any wide distribution

of pore size in the porous UF or MF membranes. For mem-

branes with only a single coating of PSSA (Figure 3(b)), the

surface roughness appears a little high, while the surface with

an added PEI coating (Figure 3(c)) appears relatively smooth.

Therefore, a double coating appears to be better than a single

coating in terms of surface morphology.

1 Permeate concentration–

Feed concentration---------------------------------------------------------------- 100×

Figure 1. Schematic diagram of the experimental apparatus for the

preparation of hollow fiber composite membranes by the PSP

method.

Figure 2. Performance test apparatus for NF hollow fiber composite

membranes.

Preparation of Hollow Fiber NF Membranes by PSP Method and Their Performance Studies 293

Polymer(Korea), Vol. 40, No. 2, 2016

Membrane Performance. First, we examined the mem-

brane performance with variations in coating concentration of

PSSA (Figure 4). PSSA of various concentrations was coated

onto the PVDF porous membrane surface under the coating

conditions of IS=0.3, coating time 2 min, and pressure 4 atm.

Then PEI 3 wt% was coated onto this layer for 1 min at IS=0.1

and the same pressure. PEI can be coated easily to form a poly-

electrolyte complex through ionic cross-linking between the

sulfonic acid group in PSSA and the amine groups in PEI. The

pressure at the second coating provides the membrane with

robustness.

As the PSSA concentration increased, the flux decreased

noticeably from 183 LMH (L/m2h) at 1 min coating to 102

LMH at 4 min. Meanwhile, rejection increased from 68% at 1

min coating to 79% at 3 min and then fell to 75% at 4 min, but

there was no remarkable change in rejection for coating with

a PSSA concentration over 2%. This absence of any differ-

ential change may be due to the ionic cross-linking between

PSSA and PEI. In case of a 1% PSSA concentration, there

may be the least number of sulfonic acid sites that can react

with the amine sites in 3% PEI. As the PSSA concentration

increases, the number of reaction sites increases, which

resulted in a salt rejection of 79% at 3% PSSA. Also the flux

decreased to 104 LMH for 3% PSSA and then no step change

was observed to 4% PSSA, although rejection was reduced

from 79.1% to 75.3%. From the viewpoint of flux and rejec-

tion, the best coating condition was with a 3% PSSA con-

centration.

Next the effect of drying time after the second coating step

was investigated (Table 1). As drying time increased, there was

no distinction in rejection while flux was reduced. From this

point, a drying time of 30 min was applied to the membrane

preparation rather than 60 min because flux at a drying time of

30 min was larger. Normally, annealing at a higher temperature

causes the membrane to become more compact, which could

be expected to reduce flux and elevate rejection (or selec-

tivity). Although our experimental results showed a reduction

in flux of more than 10%, rejection was almost unchanged.

In case of double coating, the coating time at each coating

step may have an important effect on the resulting membrane

Figure 3. SEM images of PVDF membrane surfaces coated in accordance with the conditions (a) pristine PVDF membrane; (b) PSSA 2%

for 2 min at 4 atm; (c) PEI 3% coating onto PSSA 2% coating layer for 1 min at 4 atm.

Figure 4. Effect of PSSA concentration on flux and rejection for

100 ppm NaCl (1st coating: IS = 0.3, coating time/pressure = 2 min/

4 atm, 10 min drying at 60 oC; 2nd coating: IS = 0.1, coating time/

pressure = 1 min/4 atm, 30 min drying at 60 oC).

Table 1. Effect of Drying Time at 2nd Step (PEI Coating) on

Membrane Performance

Drying time (min) Flux (LMH) Rejection (%)

30 119.3 78.9

60 105.3 78.6

*1st coating conditions: PSSA 2%, IS=0.3, coating time/pressure=2

min/4 atm, 10 min drying at 60 oC; 2nd coating conditions: PEI 3%,

IS=0.1, coating time/pressure=1 min/4 atm, drying at 60 oC.

294 S. H. Kwon and J. W. Rhim

폴리머, 제40권 제2호, 2016년

performance; therefore, the effect of the coating time at each

step was investigated (Figure 5). In general, if the coating time

increases, it is expected that flux will decrease while rejection

will increase; however, unexpectedly as shown in Figure 5, ini-

tially as the coating time increased both flux and rejection

increased. Then, as the coating time progressed, both flux and

rejection decreased as shown by the result for the PSSA 6 min/

PEI 3 min coating. From this experimental study, the optimum

coating time condition is PSSA 4 min/PEI 2 min.

We do not have any explanation for some results obtained up

to this point as follows: first, a decrease of rejection along with

a decrease of flux (Figure 4); second, a stable rejection ratio

despite a decrease in flux (Table 1); and, third, an increase of

coating time leading to an increase of both flux and rejection

followed by a further increase of coating time leading to a

decrease of flux and rejection (Figure 5). These phenomena

could not be understood in terms of changes in thickness, coat-

ing time, drying time, etc. We assume that the above men-

tioned phenomena are caused by other effects, for example, the

ionic balance between the coated layers and the feed solution

(the most probable cause). We will investigate in more detail

and prepare another paper soon.

The effect of operating time on the flux and rejection of the

composite membranes formed by the ionic bonds between

PSSA and PEI was determined (Figure 6). Flux increased at

the initial stage and then stabilized at about 80 LMH, while

rejection increased from 71% to reach 81% for an operating

time of 4 h. These results indicate an ionic balance effect

because the increases in both flux and rejection appear at sim-

ilar times. To shorten the stabilization time in both flux and

rejection, the membrane surface was cross-linked with IPC

0.1% in hexane. The stabilization time decreased as expected,

but there was a greater than expected reduction in flux. IPC

reacts with the amines in PEI; therefore, amide groups may be

formed that are not related with the ionic balance effect. Sur-

face cross-linked composite membranes may be more durable

than pristine composite membranes.

Then the cross-linking time was investigated to determine its

effect on membrane performance (Figure 7). After the cross-

linking reaction was carried out, there was a definite reduction

in flux by about half, but contrary to expectations rejection was

maintained with only slight differences. With the cross-linking

Figure 5. Effect of coating times for 1st and 2nd coating materials on

flux and rejection for 100 ppm NaCl (coating conditions: PSSA 2%/

PEI 3%, IS = 0.3/0.1, drying temp. = 60 oC, drying time = 10 min/

30 min).

Figure 6. Membrane performance by operating time for not cross-

linked and cross-linked with 0.1% IPC (coating conditions: PSSA

2%/PEI 3%, coating time and pressure = 4 min/2 min and 4 atm/4

atm, IS = 0.3/0.1, drying temp. = 60 oC, drying time = 10 min/30

min): (a) flux; (b) rejection.

Preparation of Hollow Fiber NF Membranes by PSP Method and Their Performance Studies 295

Polymer(Korea), Vol. 40, No. 2, 2016

time extended to 10 min, further reaction led to a greater

reduction in flux, but as before there was no big change in

rejection. In this case, the time effect of the cross-linking con-

cerns mainly flux changes rather than rejection changes.

Next the effect of using a cross-linking agent concentration

of 0.3% IPC instead of the previous 0.1% was investigated

(Figure 8). Rejection was elevated slightly because the

increase of the cross-linking concentration leads to more reac-

tions; however, the influence of reaction time on rejection was

relatively small. On the other hand, flux was reduced to one-

seventh after a cross-linking time of 30 sec and then by one-

third again after 3 min. From these results (Figures 7 and 8),

it is concluded that membrane performance is more affected by

the concentration of the cross-linking solution than by the

cross-linking time; therefore, we decided to examine the effect

of the concentration of the cross-linking solution in more

detail.

To examine the effect of concentration of the cross-linking

solution, membrane performance was examined at concen-

trations varying from 0.01% to 1% while the reaction time was

fixed at 30 sec (Figure 9). At each concentration, flux showed

Figure 7. Effect of cross-linking time on membrane performance

(coating conditions: PSSA 2%/PEI 3%, coating time and pressure =

4 min/2 min and 4 atm/4 atm, IS = 0.3/0.1, drying temp. and time

= 60 oC/60 oC and 10 min/30 min).

Figure 8. Membrane performance for different reaction times using

0.3% IPC as the cross-linking agent (coating conditions: PSSA 2%/

PEI 3%, coating time and pressure = 4 min/2 min and 4 atm/4 atm,

IS = 0.3/0.1, drying temp. and time = 60 oC/60 oC and 10 min/

30 min).

Figure 9. Membrane performance for different concentrations of

IPC used as the cross-linking agent (coating conditions: PSSA 2%/

PEI 3%, coating time and pressure = 4 min/2 min and 4 atm/4 atm,

IS = 0.3/0.1, drying temp. and time = 60 oC/60 oC and 10 min/30

min): (a) flux; (b) rejection.

296 S. H. Kwon and J. W. Rhim

폴리머, 제40권 제2호, 2016년

consistent values for all operating times, except with a con-

centration of 0.01%. Also flux decreased as the concentration

increased with flux reduced by almost 10 times with the 1%

concentration compared with the 0.01% concentration. On the

other hand, rejection showed a steady state with consistent val-

ues from the earliest time at low concentration ranges up to

0.3%, but at concentrations from 0.5% the time to reach the

steady state became longer and the rejection values were lower

as the concentration increased. The lowest values for both flux

and rejection at about 10 LMH and mid-70% respectively

were found with the 1% concentration.

The coating layer of the membrane used in this study is

formed by the first PSSA and the second PEI coatings and

these two polymers bond ionically with each other, i.e. as an

alternate expression, cross-linking by mutual ionic charges.

The introduction of IPC after the ionic cross-linking induces

further cross-linking between IPC and the amine groups in PEI

that remain unreacted after amine groups in PEI had reacted

with the carboxylic acid groups in PSSA. This additional

cross-linking produces a very sturdy membrane surface that

enables the feed to permeate the coated layer freely and, as a

result, the flux is very low. This phenomena appears to start

when the concentration of the cross-linking solution is 0.5% or

higher. Generally, it can be said that rejection increases if a

more compact structure is formed by cross-linking, i.e. the

sieving mechanism has a substantial effect. As a result, rejec-

tion increased at concentrations up to 0.3%, but as the con-

centration increased further there were ever greater decreases

in rejection so that rejection at 1% was the same as rejection

at 0.01%. Another unusual thing is that with an increase in

concentration from 0.5% to 1% it took longer for a steady state

flux to be reached. Probably this is due to excessive cross-link-

ing that leads to glassification of the membrane surface. Sur-

face glassification results in less flux and the excessive

modification of the surface changes the chemical properties so

that the rejection is also decreased.

Conclusions

The composite membranes by utilizing “phase separation

and pressurization” method were characterized in terms of flux

and rejection for 100 ppm NaCl at 4 atm. From this study, sev-

eral conclusions can be drawn as follows:

· Porous PVDF membranes were coated effectively as SEM

images showed the disappearance of pores from the surfaces.

· After application of the second coating, the drying time had

no effect on rejection, but flux was reduced by about 10% to

105.3 LMH with a shorter drying time of 30 min.

· The optimum individual coating times for PSSA and PEI

were 4 min and 2 min respectively with flux of 148.3 LMH

and rejection of 81.7%.

· A PSSA/PEI coated membrane cross-linked with IPC

showed no difference in rejection from a pristine PSSA/PEI

coated membrane; however, flux decreased from 148 LMH to

80 LMH.

· Cross-linking time had no effect on rejection, which

recorded 80.6% and 82.5% at 30 sec and 10 min respectively

compared with 81.7% rejection for the pristine membrane;

however, flux decreased to 82.4 and 40.4 LMH respectively

from 148.3 LMH with the pristine membrane.

· A high concentration of the cross-linking agent IPC (above

0.5%) offered no advantages because no improvement was

shown in either rejection or flux. It is recommended that the

concentration used is below 0.3%.

· Finally, as mentioned above, some results cannot be under-

stood and there may be some other effect that influences the

relationship between the feed and the membrane such as ionic

balance; therefore, this will be examined in more detail in the

next paper.

Acknowledgement: This work was supported by a Basic

Science Research Program through the National Research

Foundation of Korea (NRF) funded by the Ministry of Edu-

cation (Grant No. 2013R1A1A2008351).

References

1. X. Lu, X. Bian, and L. Shi, J. Membr. Sci., 210, 3 (2002).

2. I. Schäfer, A. G. Fane, and T. D. Waite, Water Res., 36, 1509

(2001).

3. A. Gorenflo, D. Velazquez-Padron, and F. H. Frimmel,

Desalination, 151, 253 (2002).

4. R. J. Petersen, J. Membr. Sci., 83, 81 (1993).

5. R. J. Petersen and J. E. Cadotte, Handbook of Industrial

Membrane Technology, Reprint edition, Noyes Publications,

Westwood, p 307 (1990).

6. Y. Tang, Q. S. Fu, A. P. Robertson, C. S. Criddle, and J. O.

Leckie, Environ. Sci. Technol., 40, 7343 (2006).

7. Y. Tang, Y. N. Kwon, and J. O. Leckie, J. Membr. Sci., 287, 146

(2007).

8. Y. Tang, Y. N. Kwon, and J. O. Leckie, Desalination, 242, 149

(2009).

9. C. Kim, H. G. Yoon, and K. H. Lee, J. Appl. Polym. Sci., 84, 1300

(2002).

Preparation of Hollow Fiber NF Membranes by PSP Method and Their Performance Studies 297

Polymer(Korea), Vol. 40, No. 2, 2016

10. Y. Hayakawa, N. Trasawa, E. Hayashi, and T. Abe, J. Appl.

Polym. Sci., 62, 951 (1996).

11. Z. P. Zhao, J. Li, D. Zhang, and C. X. Chen, J. Membr. Sci., 232,

1 (2004).

12. B. Bae, B. H. Chun, H. Y. Ha, I. H. Oh, and D. Kim, J. Membr.

Sci., 202, 245 (2002).

13. T. Yamaguchi, S. Nakao, and S. Kimura, Macromolecules, 24,

5522 (1991).

14. M. Ulbricht and H.-H. Schwarz, J. Membr. Sci., 136, 25 (1997).

15. A. Akbari, S. Desclaux, J. C. Rouch, and J. C. Remigy, J. Membr.

Sci., 297, 243 (2007).

16. C. Qiu, Q. T. Nguyen, and Z. Ping, J. Membr. Sci., 295, 88

(2007).

17. H. Yanagishita, J. Arai, T. Sandoh, H. Negishi, D. Kitamoto, T.

Ikegami, K. Haraya, Y. Idemoto, and N. Koura, J. Membr. Sci.,

232, 93 (2004).

18. H. Yanagishita, D. Kitamoto, K. Haraya, T. Nakane, T. Okada, H.

Matsuda, Y. Idemoto, and N. Koura, J. Membr. Sci., 188, 164

(2001).

19. A. Musale and A. Kumar, Sep. Purif. Technol., 21, 27 (2001).

20. G. Decher, Science, 277, 1232 (1997).

21. B. W. Stanton, J. J. Harris, M. D. Miller, and M. L. Bruening,

Langmuir, 19, 7038 (2003).

22. S. U. Hong and M. L. Bruening, J. Membr. Sci., 280, 1 (2006).

23. X. F. Li, S. D. Feyter, D. J. Chen, S. Aldea, P. Vandezande, F. D.

Prez, and I. F. J. Vankelecom, Chem. Mater., 20, 3876 (2008).

24. J. Miao, G. H. Chen, and C. J. Gao, Desalination, 181, 173

(2005).

25. R. H. Huang, G. H. Chen, M. K. Sun, Y. M. Hu, and C. J. Gao,

J. Membr. Sci., 286, 237 (2006).

26. T. T. Dong, G. H. Chen, and C. J. Gao, J. Membr. Sci., 304, 33

(2007).

27. R. H. Huang, G. H. Chen, M. K. Sun, and C. J. Gao,

Desalination, 239, 38 (2009).

28. T. W. Xu and W. H. Yang, J. Membr. Sci., 215, 25 (2003).

29. T. Jin, Q. F. An, Q. Zhao, J. W. Qian, and M. H. Zhu, J. Membr.

Sci., 347, 183 (2010).

30. Y. M. Guo, W. Geng, and J. Q. Sun, Langmuir, 25, 1004 (2009).

31. Y. Ji, Q. Ana, Q. Zhao, H. Chen, J. Qian, and C. Gao, J. Membr.

Sci., 357, 80 (2010).

32. J. W. Rhim, B. Lee, H. H. Park, and C. H. Seo, Macromol. Res.,

22, 361 (2014).

33. S. I. Cheong, B. Kim, H. Lee, and J. W. Rhim, Macromol. Res.,

20, 629 (2013).

34. C. J. Park, S. P. Kim, S. I. Cheong, and J. W. Rhim, Polym.

Korea, 36, 810 (2012).

35. B. Kim, H. Lee, B. Lee, S. Kim, S. I. Cheong, and J. W. Rhim,

Polym. Korea, 35, 438 (2011).

36. P. D. Ries and C. J. McDonald, WO. 003878 (1995).

37. Reverse osmosis and nanofiltration membrane module for water

supply, Korea Water and Wastewater Works Association

(KWWA), Seoul, 23 June 2009.